Dynamic flow modeling with interregional dependency effects: an application to structural change in the U.S. migration system.

Causative matrix methods can be used to project levels of population change, to monitor changing migration trends, and to aid in forecasting movement during periods of consolidation and dissipation. They are appealing because they provide measures of the changing strengths of all interregional dependency effects. Separate competing destinations and competing origins perspectives on temporal change can be obtained. The column sums and the eigenvalues provide useful aggregate gauges of the relative strengths of regional shifts. Patterns of U.S. interregional migration from 1935 to 1982 are examined using the causative matrix approach. Trends in the gross migration streams underlying the dramatic increase in core-periphery net migration taking place in the 1970s are examined, as is a more recent shift in the major source area of core region net outflow.